With the spread of digital equipment, a copy of a digital content (hereinafter referred to as a “content”) can be created easily. As a result, illegal copying which ignores a copyright is increasing, and thus protection of the copyright is becoming an important issue.
As one of solutions to the protection of the copyright, a “digital watermark” is being considered. The “digital watermark” is a technique of embedding information relating to the copyright and the like in a secret manner which cannot be perceived by humans.
Normally, various processing is applied to the content. For example, a storage media has a limitation in capacity thereof, and thus, irreversible compression of the content is performed so as to reduce capacity required for storage thereof. Further, it is assumed that processing such as format conversion and filtering is applied to the content in order to delete intentionally the information which is embedded as the digital watermark into the content.
Therefore, the digital watermark requires robustness against various types of processing as above described.
As a conventional technique to improve the tolerance of the digital watermark, for example, a technique that improves the robustness against scale-up/scale-down processing is known (e.g. see patent document 1).
FIG. 26 is a block diagram showing a conventional technique for embedding and detecting a digital watermark as described in patent document 1.
First, a conventional digital watermark embedding device embeds embedding data into an original image in accordance with equation 1 as follows.i(x,y)=.s(x,y)+(x,y)  [Equation 1]In equation 1, (x,y) indicates coordinates of a pixel, is(x,y) indicates a pixel value of the original image at the coordinates (x,y), p(x,y) indicates a pattern to be added, for embedding information, to the original image at the coordinates (x,y), and i(x,y) indicates a pixel value of a watermarked image into which the information is embedded at the coordinates (x,y).
Here, p(x,y) further satisfies a condition of equation 2 as follows.p(x+!,y)=(x,y)  [Equation 2]Note that h indicates a cycle period of the pattern.
As above described, there is a case where a scale-up or scale-down processing is applied to an embedded image obtained by embedding information to the original image. Hereinafter, a target image with respect to which a conventional digital watermark detecting device executes detection of the digital watermark is referred to as a “processing target image”.
Before detecting the digital watermark included in the processing target image, the digital watermark detecting device obtains a scale-up or scale-down ratio of the processing target image in accordance with an autocorrelation defined in equation 3. Next, based on the obtained scale-up or scale-down ratio, the digital watermark detecting device restores the embedded image by scaling up or scaling down the processing target image, and then detects watermark information.
                    Co        =                                            ∑              x                        ⁢                                          ∑                y                            ⁢                                                                    p                    ′                                    ⁡                                      (                                          x                      ,                      y                                        )                                                  ⁢                                                      p                    ′                                    ⁡                                      (                                                                  x                        +                        o                                            ,                      y                                        )                                                                                                                                            ∑                  x                                ⁢                                                      ∑                    y                                    ⁢                                                                                    p                        ′                                            ⁡                                              (                                                  x                          ,                          y                                                )                                                              2                                                                        ⁢                                                            ∑                  x                                ⁢                                                      ∑                    y                                    ⁢                                                                                    p                        ′                                            ⁡                                              (                                                                              x                            +                            o                                                    ,                          y                                                )                                                              2                                                                                                          [                  Equation          ⁢                                          ⁢          3                ]            Here, p′ (x, y) indicates an embedded pattern element extracted from a detecting image i′(x,y).
When the scale-up/scale-down ratio is represented by ρ, the autocorrelation Co has a peak at o which satisfies equation 4 as follows.ρ=/h  [Equation 4]
By searching the peak of the autocorrelation Co, the scale-up/scale-down ratio of the processing target image with respect to the embedded image can be obtained.
FIG. 27A is a schematic diagram showing a conventional scale-up/scale-down processing, and more specifically, is a diagram showing a case where the autocorrelation does not have the peak since o does not satisfy equation 4. FIG. 27B is a schematic diagram showing the conventional scale-up/scale-down processing, and more specifically, is a diagram showing a case where the autocorrelation has the peak since o satisfies equation 4.
Patent Document 1: Japanese Laid-Open Patent Publication No. 2002-111994